Fast sum and other functions in matlab?
3 vues (au cours des 30 derniers jours)
Afficher commentaires plus anciens
Hi. I have very large matrices and they cost a lot to sum up or do any other basic calculation. Can anyone please recommend faster ways of doing basic matrix operations?
Example of a few very expensive lines of code that I have (matrices img, A, and B are very big, and img is a UHD image):
A = img(s_row:end_row, s_col:end_col);
C = sum(sum(abs(A))) / max(max(abs(B)));
E(s_row:end_row, s_col:end_col) = (1 - (const1 / C)) * D .* const2;
0 commentaires
Réponses (1)
John D'Errico
le 13 Août 2016
Modifié(e) : John D'Errico
le 13 Août 2016
So, you think if they had faster ways to do something as common as a sum, they would not implement it? Nope, only slow code provided. :)
One minor enhancement. This may be slightly faster than a fragment you have written.
sum(abs(A(:)))/max(abs(B(:)))
I doubt it is much of a speed bump though.
You can look into using tools like your GPU to do some problems faster. (Which needs a separate tool.) Or, if you are doing a lot of these, the parallel computing TB may help. Speed does not come easily, and often, not cheaply.
Better is often to look more carefully at WHY you are doing those computations, and see if there is a better way.
0 commentaires
Voir également
Catégories
En savoir plus sur Resizing and Reshaping Matrices dans Help Center et File Exchange
Produits
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!